It sounds like DNews may have died during the nightly expire process (This usually occurs at 2am). The first thing to check is the amount of memory available on the syste as during the expire the memory usage will increase substancially, and can lead to symptoms similar to what you describe.
I suggest upgrading to the latest version in case this is a problem that has been addressed recently.
- Roydon L.
[EMAIL PROTECTED] wrote:
Hi all,
Our DNews server (5.5b3) running on Alpha Tru64 platform works fine for quite some time. However, I found problem this morning that the dnews process was defunct around 03:00 and server not responding. We got no way to kill the processes and had to restart the system. After restart, Dnews was not responding after startup and there was this message found in the dnews.log (last few lines and 40GB of free space in /news/spool):
17 08:12:09 0:info: Bucket check {db_38_1.itm} 17 08:12:09 0:info: Bucket check {db_127_1.itm} 17 08:12:09 0:info: Bucket check {db_1_1.itm} 17 08:12:09 0:info: Correcting free space in file from 1233355 to 1285231 17 08:12:09 0:info: Bucket check {db_32_1.itm} 17 08:12:09 0:info: Bucket check {db_17_1.itm} 17 08:12:09 0:info: Correcting free space in file from 679366 to 737448 17 08:12:09 0:info: Bucket check {db_30_6.itm} 17 08:12:09 0:error: Correcting usage for /news/spool/dnews/spool/db_30_6.itm from 15847818 to 15868398
and dfeed process did not start at all. Although the dnews process was running, it did not respond to client nor tellnews command (tellnews <cmd> just hung around). The dnews process again could not be killed. Any idea?
Thanks in advance. KC Chang
